Beamr Imaging to launch GPU-accelerated video compression solution

Beamr Imaging (BMR) will launch a high-performance, high-quality video compression solution designed for autonomous vehicles at NVIDIA GTC Paris, taking place June 10-12, 2025, as part of Viva Technology 2025, Europe’s biggest startup and tech event. In the development of autonomous driving, video is the dominant data type. A single vehicle produces terabytes of video data daily, and training a single autonomous model may require tens to hundreds of petabytes. Beamr’s proprietary technology, with a proven track record in high-efficiency video compression trusted by global media players, now addresses a pressing, costly challenge for autonomous vehicles and machine learning teams: managing video data at scale, including long-term storage and the significant infrastructure investment required. Beamr’s Content-Adaptive Bitrate technology, built on NVIDIA accelerated computing, reduces real-world autonomous driving and synthetic video file sizes by up to 50%. This is achieved while preserving visual quality and critical visual features essential for training autonomous driving models. By addressing key storage, compute, and bandwidth constraints in AI pipelines, it significantly reduces operational costs.